This function extends the classical Sobol indices to a stochastic simulator by first computing a quantity of interest (QoI) for each input point, such as the mean of repeated runs.
sobol4r_qoi_indices(
model,
X1,
X2,
qoi_fun = base::mean,
nrep = 1000,
order = 2,
nboot = 0,
type = c("soboljansen", "sobol", "sobol2007", "sobolEff", "sobolmartinez"),
...
)An object of class "sobol" with QoI-based Sobol indices.
Stochastic model function that takes a matrix or data.frame
X and returns a numeric vector of length nrow(X).
Two base designs (matrices or data.frames).
Function used to summarize the repetitions
(default is mean).
Number of repetitions of the stochastic model for each design point.
Maximum interaction order (1 or 2).
Number of bootstrap replicates for Sobol indices.
Which estimator to use. Any sensitivity Sobol helper is
supported: "sobol", "sobol2007", "soboljansen",
"sobolEff", or "sobolmartinez". Defaults to
"soboljansen", the most robust general-purpose choice.
Additional arguments passed to model.